Quote:
Originally posted by dimmreaper
If your chassis temperature and CPU temperature are increasing in a non-linear fashion, I suggest you check your testing equipment for accuracy. On-board health monitoring information is sometime "curved" intentionally, and as such is totally non-reliable. |
Very true, Jeff. I'll be a happy puppy when mainboard manufacturer's products accurately measure the internal temperature reported by AMD Morgan cored Durons and Athlon XP's. And hurry up about it! I can't remember whether AMD or the mobo maker's fumbled this. Whichever, make it work!!
I wrote that the ambient and case temperatures remain linear, it's just that ~4C is added to the case temp at a certain point. After the "bonus 4C" is present, the outside and case temperature relationship does remain linear.
As I don't plan to
OC any AMD processors, I'm not very concerned that the reported temps from the mainboard's sensors are not worth a hoot for accuracy, as you wrote. Careful installation of
HSF's that are powerful enough for "medium" overclocking on processors running at "stock" seems a wise thing to do. So I do it. This plus 100 cfm through the case assuages any concern that I'm burning up a CPU--but don't know it because the reported temps are lies.
And I also remove the mainboard from the case to install new
HSF's. Because I can not eyeball the CPU/
HSF junction without doing so. This has been known to delay the installation of new
HSF's. Kinda.